This competing renewal of R01 MH58076 proposes a multi-center, randomized clinical trial to evaluate the efficacy of a CNS-targeted antiretroviral treatment (ART) strategy for individuals with HIV-associated neurocognitive impairment (HNCI). Although highly active antiretroviral therapy (HAART) has reduced the incidence of severe dementia and may improve neurocognitive function, individual patient responses are variable, and the prevalence of HNCI remains high. Because HNCI is a significant burden to persons living with HIV infection, caregivers, and the healthcare system, and because it often persists despite HAART, the development of effective treatment strategies is of great public health importance. Consensus ART guidelines provide no specific recommendations on the management of HNCI. Evidence from in vitro and in vivo experiments and uncontrolled human studies has failed to resolve controversy on whether antiretroviral drug (ARV) CNS penetration and cerebrospinal fluid (CSF) virologic suppression are clinically important. Finally, no randomized, controlled trials have compared the neurocognitive benefits of regimens with differing CNS penetration. The chemical profile of many ARVs currently in development suggests that their CNS penetration will be negligible. Thus data on CNS penetration will become increasingly important as new agents are added into the clinical armamentarium. The proposed trial design is based on findings from the current funding period of this grant. Briefly, among individuals with HNCI initiating new ART, those on more CNS penetrating regimens had greater CSF viral load (VL) reduction, and neurocognitive outcomes were better in those who achieved CSF virologic suppression. Patients eligible for the study will: 1) meet consensus diagnostic criteria for HNCI;and 2) be anticipating initiation of a new ART regimen. A total of 120 patients at 3 sites will be randomized 1:1 to receive a CNS-targeted (CNS-T) ART strategy versus a non-CNS-T (comparison) strategy. The primary outcome, an assessment of change in global neurocognitive performance on a reliable and well-validated test battery, will be evaluated at 16 weeks, with continued follow-up to 32 weeks. Secondary aims will be to compare CNS-T and non-CNS-T ART on measures of CSF and plasma HIV suppression, and to delineate the impact of CNS-T on specific domains of neurocognitive functioning and on activities of daily living. Additional assessments that address explanatory goals will include ARV pharmacokinetics in plasma and CSF, and ARV resistance assays on CSF viral isolates. Samples will be stored for future use in studies aimed to evaluate viral fitness, neuropathogenicity and compartmental evolution. The proposed clinical trial would provide the level of evidence needed to formulate ART guidelines specific to HNCI.